A Hierarchical Unsupervised Spectral Clustering Scheme for Detection of Prostate Cancer from Magnetic Resonance Spectroscopy (MRS)

  • Pallavi Tiwari
  • Anant Madabhushi
  • Mark Rosen
Part of the Lecture Notes in Computer Science book series (LNCS, volume 4792)

Abstract

Magnetic Resonance Spectroscopy (MRS) along with MRI has emerged as a promising tool in diagnosis and potentially screening for prostate cancer. Surprisingly little work, however, has been done in the area of automated quantitative analysis of MRS data for identifying likely cancerous areas in the prostate. In this paper we present a novel approach that integrates a manifold learning scheme (spectral clustering) with an unsupervised hierarchical clustering algorithm to identify spectra corresponding to cancer on prostate MRS. Ground truth location for cancer on prostate was determined from the sextant location and maximum size of cancer available from the ACRIN database, from where a total of 14 MRS studies were obtained. The high dimensional information in the MR spectra is non linearly transformed to a low dimensional embedding space and via repeated clustering of the voxels in this space, non informative spectra are eliminated and only informative spectra retained. Our scheme successfully identified MRS cancer voxels with sensitivity of 77.8%, false positive rate of 28.92%, and false negative rate of 20.88% on a total of 14 prostate MRS studies. Qualitative results seem to suggest that our method has higher specificity compared to a popular scheme, z-score, routinely used for analysis of MRS data.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Catalona, W., et al.: Measurement of Prostate-Specific Antigen in serum as a Screening Test for Prostate Cancer. J. Med. 324(17), 1156–1161 (1991)Google Scholar
  2. 2.
    Kurhanewicz, J., et al.: Analysis of a Complex of Statistical Variables into Principal Components. Magnetic Resonance in Medicine 50, 944–954 (2003)CrossRefGoogle Scholar
  3. 3.
    McKnight, T., et al.: An Automated Technique for the Quantitative Assessment of 3D-MRSI Data from Patients with Glioma. Journal of Magnetic Resonance Imaging 13, 167–177 (2001)CrossRefGoogle Scholar
  4. 4.
    Pels, P., Ozturk-Isik, et al.: Quantification of Prostate MRSI Data by Model-Based Time Domain fitting and Frequency Domain Analysis. NMR in Biomedicine 19, 188–197 (2006)CrossRefGoogle Scholar
  5. 5.
    Ma, J., Sun, J.: MRS Classification based on Independent Component Analysis and Support Vector Machines. In: IEEE Intl. Conf. on Hybrid Intel. Syst., pp. 81–84. IEEE Computer Society Press, Los Alamitos (2005)Google Scholar
  6. 6.
    Simonetti, A., Melssen, W., Edelenyi, F., et al.: Combination of Feature-Reduced MR Spectroscopic and MR Imaging Data for Improved Brain Tumor Classification. NMR in Biomedicine 18, 34–43 (2005)CrossRefGoogle Scholar
  7. 7.
    Lee, G., Madabhushi, A., Rodriguez, C.: An Empirical Comparison of Dimensionality Reduction Methods for Classifying Gene and Protein Expression Datasets. ISBRA (2007)Google Scholar
  8. 8.
    Anant Madabhushi, J.S., et al.: Graph Embedding to Improved Supervised Classification and Novel Class Detection: Application to Prostate Cancer. In: Duncan, J.S., Gerig, G. (eds.) MICCAI 2005. LNCS, vol. 3749, pp. 729–737. Springer, Heidelberg (2005)CrossRefGoogle Scholar
  9. 9.
    Madabhushi, A., Feldman, M., et al.: Automated Detection of Prostatic Adenocarcinoma from High-Resolution Ex Vivo MRI. IEEE Transactions on Medical Imaging 24(12) (2005)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2007

Authors and Affiliations

  • Pallavi Tiwari
    • 1
  • Anant Madabhushi
    • 1
  • Mark Rosen
    • 2
  1. 1.Department of Biomedical Engineering, Rutgers UniversityUSA
  2. 2.Department of Surgical Pathology, University of PennsylvaniaUSA

Personalised recommendations